agentleFS
NoorQureshiNoorQureshi/SploitAgent/skills/code-review/code-review-python/SKILL.md

code-review-python

Skill26 starsChanged 31 days agoMITView on GitHub

Security review of Python code — dangerous sinks and framework-specific pitfalls (Django/Flask/ FastAPI). Load when reviewing a Python codebase/PR, on .py source in scope, or "review this Python". Signals: requirements.txt/pyproject, Django/Flask/FastAPI, ORMs, pickle/yaml, subprocess.

  • Reads credentials

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Files
SKILL.md

Python security code review

When it applies

You're reading Python source (a repo, a PR, a service). This is the language-specific companion to code-review-methodology — the exact sinks and framework gotchas to grep and trace.

Why it works

Most severe Python bugs come from a known set of sinks plus framework misuse. Grep the sinks, trace each argument to a user source, and check the framework's safe-vs-unsafe API was used.

Sinks & patterns (grep, then trace to user input)

  • Command exec: os.system, subprocess.*(..., shell=True), os.popen — shell=True + user input = injection.
  • Code eval: eval, exec, pickle.loads, yaml.load (without SafeLoader), marshal — deserialization/eval RCE.
  • SQL: raw/f-string queries, .raw(), .extra(), cursor.execute("... %s" % x) — use params, not formatting.
  • SSRF: requests.get/urllib/httpx on a user URL (→ web-ssrf).
  • Path/upload: open/send_file/os.path.join with user paths (traversal); zip extraction (zip-slip).
  • Template (SSTI): render_template_string, Jinja from user input (→ web-ssti).
  • Secrets: hardcoded keys/passwords; DEBUG=True in prod.

Framework specifics

  • Django: mark_safe/|safe (XSS), .raw()/.extra() (SQLi), DEBUG=True (info leak), SECRET_KEY exposure (session forgery), missing @login_required/object-level checks (IDOR), ALLOWED_HOSTS='*', pickle session serializer.
  • Flask: render_template_string (SSTI), debug=True (Werkzeug console RCE), weak SECRET_KEY (session tampering — flask-unsign), send_file traversal.
  • FastAPI: missing dependency-injected auth on routes, over-broad CORS, Pydantic not enforcing server-side authz (BOLA at the data layer).

Method

rg -n "shell=True|eval\(|exec\(|pickle.loads|yaml.load\(|render_template_string|\.raw\(|\.extra\("; run bandit -r . and semgrep --config auto; triage by exploitable source→sink.

Gotchas

  • yaml.safe_load and parameterized ORM queries are the safe variants — confirm which is used.
  • Bandit is noisy; rank by user-controllable input reaching the sink.

References

Bandit; Semgrep Python rules; OWASP Django/Flask cheat sheets.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.